A computational approach to unraveling TLR signaling in murine mammary carcinoma.

We developed an agent-based model to simulate a signaling cascade which allowed us to focus on the behavior of each class of agents independently of the other classes except when they were in physical contact. A critical piece was the ratio of the populations of agents that interact with one another, not their absolute values. This ratio reflects the effects of the density of each agent in the biological cascade as well as their size and velocity. Although the system can be used for any signaling cascade in any cell type, to validate the system we modeled Toll-like receptor (TLR) signaling in two very different types of cells; tumor cells and white blood cells. The iterative process of using experimental data to improve a computational model, and using predictions from the model to design additional experiments strengthened our understanding of how TLR signaling differs between normal white blood cells and tumor cells. The model and experimental data showed that some of the differences between the tumor cells and normal white blood cells were related to NFκB and TAB3 levels, and also suggested that tumor cells lacked IRAKM-dependent feedback inhibition as a negative regulator of TLR signaling. Finally, we found that these different cell types had distinctly different responses when exposed to two signals indicating that a more biologically relevant model and experimental system should address activation of multiple interconnected signaling cascades, the complexity of which further reinforces the need for a combined computational and molecular approach.